Re: Running Grails application which requires ~3GB on the heap - possible in heroku?
It's not possible, sorry. A single dyno is limited to 512 MB of RAM (see the Acceptable Use Policy: http://policy.heroku.com/aup). And cranking up the dynos won't help, because their RAM won't pool (and Heroku goes out of its way to split your dynos across different machines anyway). On Friday, April 20, 2012 7:46:40 AM UTC-7, Chris wrote: Hi, I'm slightly confused as to whether it is possible to run an application on Heroku that requires around 3GB on the heap. This does seem a lot, but it's not a memory leak - it is just that it is necessary to have around 3,000,000 objects in memory at a certain time (yes, I will be reviewing just how 'necessary' it is, and what alternatives there are, but for now lets say it's necessary). So, this long running task would be handled by a Worker Dyno, rather than a Web Dyno? Are the limitations/capabilities of the Worker Dynos the same as the Web Dynos? And by cranking up the number of Dynos, can I run my application that will require say 3GB for one process, or is this just not possible? Sorry for the noob questions ... but I am a noob, and I've seen conflicting information, so want to get the facts straight. Cheers, Chris -- You received this message because you are subscribed to the Google Groups Heroku group. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/heroku?hl=en_US?hl=en
Config var unavailability during asset precompilation.
I'm using Sequel with Rails 3.1.0, and connecting to my database by simply passing a database url in a config var to Sequel.connect - the standard way of doing it. It seems that when Heroku runs the asset precompilation rake task during a deploy, my config vars apparently aren't available to the app as it's initializing. Without a database to connect to, it crashes. I could simply hack my way around this by rescuing from the connection error, but I think that's a bad idea, since my app relies on the database for all of its functionality - if the database is unavailable, the app SHOULD be crashing when it's trying to establish its connection. I think it makes more sense for my config vars to be available to the app during asset compilation, but I'm not sure why they're not already. Can someone from Heroku shed some light on this? Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Config var unavailability during asset precompilation.
I wound up solving this by establishing the DB connection unless Rails.env.production? Rails.groups.include?('assets'). It seems to work alright. On Sep 13, 12:22 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I'm using Sequel with Rails 3.1.0, and connecting to my database by simply passing a database url in a config var to Sequel.connect - the standard way of doing it. It seems that when Heroku runs the asset precompilation rake task during a deploy, my config vars apparently aren't available to the app as it's initializing. Without a database to connect to, it crashes. I could simply hack my way around this by rescuing from the connection error, but I think that's a bad idea, since my app relies on the database for all of its functionality - if the database is unavailable, the app SHOULD be crashing when it's trying to establish its connection. I think it makes more sense for my config vars to be available to the app during asset compilation, but I'm not sure why they're not already. Can someone from Heroku shed some light on this? Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Gzip not available on Cedar?
No problem. The issue with HTML pages not being Gzipped resolved itself somehow - now everything's being compressed just as it should be. The downside is that it ties up the dyno, and adds about 5-10 ms of processing time to my heavier pages. It would be great if Heroku could get Nginx handling this again, but oh well. Now I'm able to run three Rails instances on each dyno (via Unicorn) instead of one, so the tradeoff is worth it. Also, I had to rig up a middleware to set 'Cache-Control' headers on my assets, since that isn't done for you anymore. The two apps I've switched over are the ones that don't really use HTTP caching aside from the assets, so losing Varnish isn't a big deal. I have a third that uses HTTP caching extensively, and I'm unsure about it. Switching to Rack::Cache might be worth it for the ability to expire pages from within the app, but I'll have to look into it some more. On Jun 3, 10:56 am, John McCaffrey john.mccaff...@gmail.com wrote: Thanks for the play by play! I didn't get a chance to try out any of my apps on it yet, but I appreciate the heads up on some of the things I might run into. At first the lack of varnish seemed like a big loss, and that it would now mean that your app server is requests that used to be served by varnish. I read through this post on rack cachehttp://www.saturnflyer.com/blog/jim/2010/06/24/rack-cache-on-heroku-w..., which might not be the most up to date, but it eased my concern somewhat. Getting the official word on the gzip thing would be nice. thanks -John On Thu, Jun 2, 2011 at 8:50 PM, Chris Hanks christopher.m.ha...@gmail.comwrote: Nevermind, I found Rack::Deflater, and it seems to be Gzipping most things alright (it takes care of the assets, but not HTML, and i'm not sure why). On Jun 2, 6:30 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I tried moving an app over to Cedar, and now that I'm looking at it in Firebug it appears that Heroku isn't gzipping responses anymore. I knew that I'd have to handle my own http caching since this stack doesn't use Varnish, but I thought that Gzip was handled by Nginx and would still be available. Is this another thing we have to handle at the app level now? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- -John -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Gzip not available on Cedar?
I tried moving an app over to Cedar, and now that I'm looking at it in Firebug it appears that Heroku isn't gzipping responses anymore. I knew that I'd have to handle my own http caching since this stack doesn't use Varnish, but I thought that Gzip was handled by Nginx and would still be available. Is this another thing we have to handle at the app level now? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Gzip not available on Cedar?
Nevermind, I found Rack::Deflater, and it seems to be Gzipping most things alright (it takes care of the assets, but not HTML, and i'm not sure why). On Jun 2, 6:30 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I tried moving an app over to Cedar, and now that I'm looking at it in Firebug it appears that Heroku isn't gzipping responses anymore. I knew that I'd have to handle my own http caching since this stack doesn't use Varnish, but I thought that Gzip was handled by Nginx and would still be available. Is this another thing we have to handle at the app level now? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: varnish not caching?
I'm seeing the same behavior as craayzie. Redbot shows headers that look correct: HTTP/1.1 200 OK Server: nginx/0.7.67 Date: Wed, 01 Jun 2011 20:16:01 GMT Content-Type: application/json; charset=utf-8 Connection: keep-alive Last-Modified: Wed, 11 May 2011 19:35:02 GMT Cache-Control: max-age=86400, public X-UA-Compatible: IE=Edge,chrome=1 X-Runtime: 0.006309 Content-Length: 403 X-Varnish: 1696646971 Age: 0 Via: 1.1 varnish If I point Redbot to the same url repeatedly, it shows the same headers every time, but the logs show that each request is hitting my rails app. Varnish isn't serving it for some reason. I'm on Bamboo, so Varnish should be working, it's just... not. This app is only at one dyno currently. On May 31, 5:25 pm, craayzie flesh...@gmail.com wrote: I'm setting a 'Cache-Control: public, max-age=300' response header per the Heroku docs but when I request the resource, Heroku logs show my app processing the request and subsequent requests for the same resource continue to hit my app. Any guidance on how to troubleshoot this further to figure out why Varnish isn't caching? Here are the response headers from Heroku: HTTP/1.1 200 OK Server: nginx/0.7.67 Date: Wed, 01 Jun 2011 00:24:28 GMT Content-Type: text/html; charset=utf-8 Connection: keep-alive ETag: 8b8a779a2522e7a42fc663b76a088763 X-UA-Compatible: IE=Edge,chrome=1 X-Runtime: 0.047719 Cache-Control: public, max-age=300 Content-Length: 6921 X-Varnish: 1131028875 Age: 0 Via: 1.1 varnish -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: varnish not caching?
Huh. I remembered seeing somebody from Heroku say that they had their Varnish servers set up in a hash ring, which I thought would mean this kind of effect wouldn't happen, but I guess it does. It looks like they have 10 Varnish servers going - if I use Apache Benchmark to hit my server with 1000 simultaneous requests it hits my Rails app 10 times before everything starts being served from Varnish. Anyway, thanks for looking into it, Craayzie. On Jun 1, 7:24 pm, craayzie flesh...@gmail.com wrote: So after hearing back from support it looks like Varnish is caching just fine it's just that they have a bunch of Varnish servers and the chances of you hitting the same one twice is pretty low. Try firing off 10 simultaneous requests and you'll see some of them served from Varnish cache. Nice :) -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Changing static asset cache time
Yeah, Heroku should make this a lot easier. I experimented with one approach to this a while ago, and it seemed to work fine on my staging dyno. I haven't gotten around to actually running it in production yet, though, so use at your own risk: https://gist.github.com/998124 Let me know how it goes! On May 29, 1:26 pm, S Wrobel swro...@gmail.com wrote: I'd like my static assets to cache for longer. I think Heroku's default is 12 hours, but considering that Rails adds the ?33253523 stuff to them to indicate when a new version has been deployed, there's no reason not to cache them for months if not years. 12 hours seems way too short for me. Does anyone have any feedback on whether this is wise to do, and also how to override it? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Rails 3.1 asset pipeline causing 503?
I get this same error frequently on my development machine (Ruby 1.9.2p180 on Ubuntu Natty), so it's not Heroku-specific. On May 20, 8:01 am, David datab...@gmail.com wrote: Klaus Harti posted this message originally on May 11th and I have it in my digest list - I just can't seem to find it on Google Groups. Anyhow, I'm seeing a similar problem but with this on serving the resulting CSS: !! Unexpected error while processing request: deadlock; recursive locking Log: 2011-05-20T08:31:13+00:00 heroku[web.1]: Starting process with command: `thin -p 49201 -e production -R /home/heroku_rack/heroku.ru start` 2011-05-20T08:31:17+00:00 app[web.1]: Starting the New Relic Agent. 2011-05-20T08:31:17+00:00 app[web.1]: Installed New Relic Browser Monitoring middleware 2011-05-20T08:31:19+00:00 app[web.1]: Thin web server (v1.2.6 codename Crazy Delicious) 2011-05-20T08:31:19+00:00 app[web.1]: Maximum connections set to 1024 2011-05-20T08:31:19+00:00 app[web.1]: Listening on 0.0.0.0:49201, CTRL+C to stop 2011-05-20T01:31:20-07:00 heroku[web.1]: State changed from starting to up 2011-05-20T08:31:27+00:00 app[web.1]: 2011-05-20T08:31:27+00:00 app[web.1]: 2011-05-20T08:31:27+00:00 app[web.1]: Started GET / for 72.229.127.215 at 2011-05-20 01:31:27 -0700 2011-05-20T08:31:27+00:00 app[web.1]: Processing by HomeController#index as HTML 2011-05-20T08:31:27+00:00 app[web.1]: Rendered home/index.html.erb within layouts/application (15.7ms) 2011-05-20T08:31:27+00:00 app[web.1]: [Sprockets] application.css building 2011-05-20T08:31:27+00:00 app[web.1]: [Sprockets] style.css building 2011-05-20T08:31:27+00:00 app[web.1]: [Sprockets] application.js building 2011-05-20T08:31:27+00:00 app[web.1]: [Sprockets] jquery.js building 2011-05-20T08:31:27+00:00 app[web.1]: [Sprockets] jquery_ujs.js building 2011-05-20T08:31:28+00:00 app[web.1]: Rendered shared/ _tracking_footer.html.erb (0.4ms) 2011-05-20T08:31:28+00:00 app[web.1]: Completed 200 OK in 1417ms (Views: 1417.0ms | ActiveRecord: 0.0ms) 2011-05-20T08:31:28+00:00 heroku[router]: GET mysite.heroku.com/ dyno=web.1 queue=0 wait=0ms service=1469ms bytes=1834 2011-05-20T08:31:28+00:00 app[web.1]: cache: [GET /] miss 2011-05-20T08:31:28+00:00 app[web.1]: Connected to NewRelic Service at collector-1.newrelic.com:80 2011-05-20T08:31:29+00:00 app[web.1]: !! Unexpected error while processing request: deadlock; recursive locking 2011-05-20T08:31:29+00:00 heroku[router]: Error H13 (Connection closed without response) - GET mysite.heroku.com/assets/application- b1637fec79c8bbd5d3a18e84f3da878f.css dyno=web.1 queue=0 wait=0ms service=0ms bytes=0 2011-05-20T01:31:29-07:00 heroku[nginx]: GET / HTTP/1.1 | 72.229.127.215 | 1480 | http | 200 both edge and beta1 have the same results: Gemfile: source 'http://rubygems.org' #gem 'rails', '3.1.0.beta1' gem 'rails', git: 'git://github.com/rails/rails.git' gem 'pg' gem 'therubyracer-heroku', '0.8.1.pre3' # Asset template engines gem 'sass' gem 'coffee-script' gem 'uglifier' gem 'jquery-rails' Previous post: Rails 3.1 asset pipeline causing 503? Klaus Hartl klaus.ha...@googlemail.com May 11 06:06AM -0700 ^ Hello there, I was curious and tried if I can use the new assets pipeline that comes with Rails 3.1 beta (sprockets). From time to time I am getting a 503 for requests of those merged application.js/.css files, an example from my log: 2011-05-11T12:26:10+00:00 app[web.1]: cache: [GET /assets/ application-140d82c5f1fa592792bc01518bd2906e.js] fresh 2011-05-11T12:26:10+00:00 heroku[router]: GET tweektv.heroku.com/ assets/application-140d82c5f1fa592792bc01518bd2906e.js dyno=web.1 queue=0 wait=0ms service=2ms bytes=8700 2011-05-11T05:26:10-07:00 heroku[nginx]: GET /assets/ application-140d82c5f1fa592792bc01518bd2906e.js HTTP/1.1 | 84.191.168.124 | 1325 | http | 503 I am fully aware that this is beta and I did not necessarily expect it to work just so. I'm wondering if this is something related to the app or if this is on Heroku's side anyway (in other words, should I just wait for Rails 3.1 out of beta or not)... The log message isn't very helpful to me, as is what I get in the response of such 503 request (stripped html): Application Error An error occurred in the application and your page could not be served. Please try again in a few moments. If you are the application owner, check your logs for details. Last not least I have installed the Exceptional Premium add-on and I'm not getting any error notifications in there. Any insight is very much appreciated... --Klaus -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options,
Re: what external executables are available in the heroku stack?
It would also be good if our apps had access to a node.js executable, since execjs would be able to use it to compile coffeescripts in Rails 3.1 when it's released. On May 4, 9:23 am, fearless_fool rdp...@gmail.com wrote: I've dug around for a bit and haven't found a list of all the executables that are available from within a Heroku-hosted Rails app. But the list must exist -- can someone point me to it? TIA. -- ff (I actually want to know if it's possible for my Heroku hosted app to call gnuplot -- I'm guessing no, but it would be nice to know what resources are available before I write more Ruby code.) -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Cron: Run a cron job only on Monday
Try: Time.now.monday? On Mar 31, 8:16 am, Doug Naegele dougnaeg...@gmail.com wrote: Can someone help me configure Heroku Cron to only run on a certain day? So, imagine I send an email report every Monday morning. How do I write the syntax for that? Something like this: This one works: if Time.now.hour == 7 ##Sends report @7am, PST. puts Send report uri = URI.parse('http://myapp.heroku.com/messages/send_report') Net::HTTP.get(uri) puts Report is done. end This one does not work: if Time.now == Monday ##Sends report at the daily Heroku cron run, but only on Monday. puts Send report uri = URI.parse('http://myapp.heroku.com/messages/send_report') Net::HTTP.get(uri) puts Report is done. end thanks in advance -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Cron: Run a cron job only on Monday
Oh, you must be on Ruby 1.8.7, then? You'll have Time#thursday? (and the other days of the week) on 1.9, or if you're using Rails. Sorry for the confusion. Jeff's suggestion is best, but when you upgrade to 1.9 I'd suggest using Time#thursday?, since it's so much more readable. On Mar 31, 10:14 am, Doug Naegele dougnaeg...@gmail.com wrote: Thanks Chris. Sorry, it didn't work. Here's my code: if Time.now.thursday? puts Send report uri = URI.parse('http://myapp.heroku.com/messages/send_report') Net::HTTP.get(uri) puts Report put is done. end Here's the error output (after I ran heroku rake cron --trace) ** Invoke cron (first_time) ** Invoke environment (first_time) ** Execute environment ** Execute cron rake aborted! undefined method `thursday?' for Thu Mar 31 10:06:51 -0700 2011:Time /app/lib/tasks/cron.rake:28 (if Time.now.thursday? is line 28) - the following also throws an error: if Time.now == thursday . . . . (output below) rake aborted! undefined local variable or method `thursday' for main:Object this also throws an error: if Time.now.day == thursday . . . . (output below) rake aborted! undefined local variable or method `thursday' for main:Object On Mar 31, 12:32 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: Try: Time.now.monday? On Mar 31, 8:16 am, Doug Naegele dougnaeg...@gmail.com wrote: Can someone help me configure Heroku Cron to only run on a certain day? So, imagine I send an email report every Monday morning. How do I write the syntax for that? Something like this: This one works: if Time.now.hour == 7 ##Sends report @7am, PST. puts Send report uri = URI.parse('http://myapp.heroku.com/messages/send_report') Net::HTTP.get(uri) puts Report is done. end This one does not work: if Time.now == Monday ##Sends report at the daily Heroku cron run, but only on Monday. puts Send report uri = URI.parse('http://myapp.heroku.com/messages/send_report') Net::HTTP.get(uri) puts Report is done. end thanks in advance -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Action Caching... force it to use Varnish not Memcache
Not sure what you mean by this. Varnish is an HTTP cache, it sits in front of your dyno and only caches the responses that it sends - you can't use it as a key-value store like memcached. When a request is handled by varnish and doesn't even hit your app, that IS page caching. If you could write a bit more about what you're trying to do, we can offer more guidance. On Feb 19, 7:45 pm, railsnerd rails.n...@gmail.com wrote: Hi everyone Is there a way to force Action Caching to cache to Varnish and not use Memcache? Yes I fully realise this might cause (?) the next request to hit Varnish and not my app, which defeats the purpose of Action Caching and I should use Page Caching... but I have my reasons :) If there was a hack or a tweak, or any ideas, I'd be super happy! thanks -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Price reduction in custom SSL?
I know that the custom SSL plan costs $100 because it requires a dedicated EC2 instance, and the small instances cost about $60 a month. But since EC2 has now introduced micro instances, for a little under $15 a month, will we be seeing a price reduction for the custom SSL plan? Chris -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Question on shutting down Workers
Hi - I'm rigging up my own auto-scaling solution for workers. My plan to shut a worker down when it's not needed is to set $exit = true when there are no jobs left for it to do. As you can see (https:// github.com/collectiveidea/delayed_job/blob/v2.1.3/lib/delayed/ worker.rb#L74-92) this will simply end the rake jobs:work task just as if I had (for example) hit Ctrl+C. My question is, will Heroku recognize that the worker process has ended and stop charging my account when this happens? Or do I have to actually hit the Heroku API to shut it down? I'd prefer to use $exit if possible, since that will ensure that the current worker (the one with no jobs available) shuts down, whereas if I tell the Heroku API to decrement my worker count by 1, and I have multiple workers running, it could kill one of the other ones mid-job. Thanks! Chris -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Question on shutting down Workers
Alright, that makes sense. Thanks, Adam. On Jan 25, 11:12 am, Adam Wiggins a...@heroku.com wrote: I understand what you're trying to do, but this won't work. Heroku will try to keep your workers alive no matter what; when it exits it will change state to crashed and try to restart once, then again every ten minutes. The full crash policy is described here: http://docs.heroku.com/ps#crashed-process-restarts If you want to change the number of running workers, you need to use the Heroku API. Adam -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: My (flawed?) attempt to add or subtract workers via Heroku API.
You can start up as many workers (or dynos, for that matter) as you want through the command line. 24 is just how high the slider goes on the pricing page. On Jan 21, 7:52 am, rubynoob mysmilecent...@gmail.com wrote: I may misunderstand how workers get charged on heroku, but from what I've seen athttp://docs.heroku.com/background-jobsandhttp://docs.heroku.com/delayed-job, workers get charged $0.05/hr each no matter how many are running, pro-rated to the second. The maximum workers per account seems to be 24 (that's where the slider stops on their Resources page). The jobs being delayed won't be created faster than one every 30 seconds, so I assumed the first worker would spin up and grab the first job, then when the second job gets queued, a second worker would start, grabbing the second job, and so on. Each job would process in it's own worker, which would then get shut down when the job completes. One worker running three jobs that take a total of fifteen minutes to process should get charged the same as three workers running one job each for five minutes. If I'm mistaken, let me know. Thanks, Jim On Jan 20, 4:57 pm, Keenan Brock kee...@thebrocks.net wrote: Also a thought. You will need to introduce a lag when you are determining if you need more or less delayed job workers. Otherwise you will spin up too many DJs too quickly. And add/remove them very often. Incurring extra charges. Smugmug spoke about this when they were talking about their on demand photo processors a few years back. --Keenan On Jan 20, 2011, at 6:06 PM, Pedro Belo pe...@heroku.com wrote: That was a good call, you definitely don't want to store variables in config vars. Save if for constants (passwords, urls, etc). It seems like you might be getting an error due to different versions of RestClient, not sure though. What version are you using? What's the stack trace for the exception? On a side note, if it helps you can call heroku workers passing relative values, like +3, -1, etc. On Thu, Jan 20, 2011 at 12:00 PM, rubynoob mysmilecent...@gmail.com wrote: Instead of storing the count of active workers as a heroku config variable, I decided to create a table in our database to store the value in. So now I've got the problem narrowed down to the last line in the method. Here's the block of code I've now got: add_heroku_worker heroku = Heroku::Client.new(ENV['HEROKU_USERNAME'], ENV['HEROKU_PASSWORD']) myapp = heroku.config_vars(ENV['HEROKU_APP'])[HEROKU_APP] worker_count = WorkerCount.find(1) # now I'm storing the current number of active workers in a table that will always only have one record. workers = worker_count.workers qty = workers + 1 worker_count.workers = qty worker_count.save heroku.set_workers(myapp, qty) end In the heroku console, this runs smoothly until I try the last line, to which I get this error: TypeError: can't convert RestClient::Payload::UrlEncoded into String This line is formatted the same as LostBoy's workless gem, the autoscaling tree of delayed_job, and Heroku-Delayed-Job-Autoscale. I must be missing something obvious (typical newbie, huh?) ;) Thanks again for any help, Jim On Jan 20, 9:16 am, Peter Haza peter.h...@gmail.com wrote: I've done autoscaling of workers here:https://github.com/phaza/Heroku-Delayed-Job-Autoscale It's actually more like auto-shutdown of a single workers, but it works well in our environment. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Running a daemon process on Heroku
Well, since you pay for workers by the hour, I'm sure Heroku would be perfectly happy with you keeping many of them busy for a very long time. On Dec 7, 10:44 pm, Jonas jo...@jonasbnielsen.dk wrote: Nice. So that was actually what I initially thought. And there is no limitation on what that worker does? I mean, it's okay to preoccupy that worker till eternity? :) On 7 Dec., 20:53, David Dollar da...@heroku.com wrote: A Heroku worker is simply running rake jobs:work on your app so whatever happens behind that rake task is up to your app. On Dec 7, 2010, at 10:47 AM, Jonas jo...@jonasbnielsen.dk wrote: Hi guys, Hi guys, Ok, so I've succesfully developed a daemon that fires up EventMachine, subscribe to a data feed and communicate that data to Pusherapp.com Everything is working beautifully in production on my OSX development machine, and now I want to deploy to Heroku. My initial understanding was that the Heroku workers would suite this purpose well, however I've come to realize that I might be wrong. My question is, how do I run a daemon in the Heroku environment? Some facts: Ruby 1.9.2 bamboo-mri-1.9.2 (beta) Rails3 Daemons gem:https://rubygems.org/gems/daemons The deamon_generator plugin:https://github.com/dougal/daemon_generator An old Railscast that describes approx. how a daemon like this is set up:http://railscasts.com/episodes/129-custom-daemon How the daemon works: I've setup a rake task to start the daemon, briefly, this is the process: 1. Rake task calls lib/daemons/my_daemon_ctl start 2. Require some gems and start daemon: Daemons.run File.dirname(__FILE__) + /my_daemon.rb, options 3. Loads rails env and starts the EventMachine reactor: EventMachine::run { data feed and pusherapp black magic here } -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
How to determine if my app is being run on a dyno or a worker?
I have a few delayed_jobs that require data to be loaded from the database and kept in memory. I've set up an initializer to load that data (this is a Rails 3 app), but I'm not sure how to only run it on a worker process and not my dynos. I noticed, while tooling around in `heroku console`, that ENV[HEROKU_TYPE] == Dyno - I was hoping that in a worker process it would be Worker, but that doesn't seem to be working, and I don't know how to inspect the ENV hash from a worker instance. Any pointers? Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: compass causing file writing error?
In my compass initializer file I include: Sass::Plugin.options[:never_update] = true if Rails.env.production? On Nov 16, 9:57 pm, Josh Coffman joshcoff...@gmail.com wrote: I'm using compass in one of my rails apps on heroku, and every time I push it, it fails to load on the first request and shows an error page. It works fine on subsequent requests. I think compass is trying to write to public/stylesheets/ie.css. here is the error from the heroku logs: Started GET / for 72.208.78.46 at 2010-11-16 21:51:09 -0800 Errno::EROFS (Read-only file system - /disk1/home/slugs/274925_210ef9f_8bf2-58f50fc2-dbf0-4734-8c90-e918608ebf20/ mnt/public/stylesheets/ie.css - Heroku has a read-only filesystem. Seehttp://docs.heroku.com/constraints#read-only-filesystem): Anyone know how to stop it from writing? I tried using the line in production.rb to turn off Sass, but that just gives be an error about module/namespace not found. TIA, Josh -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: I added BUNDLE_WITHOUT=development test, but Heroku still installs everything
Yes, this started happening for me on Friday (worked fine up until then). and it looks like it's still going on. When I push I see Using --without development test, but it tries to install all the development/test gems anyway (which it can't, for me, because Heroku can't handle ruby-debug19). I haven't had a burning need to deploy recently, so I've been waiting for Heroku to sort it out, but it looks like they haven't, yet. Anyway, it's definitely a bug, so why not open a support ticket about it? And maybe let us know how it works out? On Nov 6, 3:06 pm, Volkan Unsal spockspla...@gmail.com wrote: I wanted Heroku's bundler to ignore some gems I have installed. It doesn't manage to do that, and as a result I am getting lots of errors. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: I added BUNDLE_WITHOUT=development test, but Heroku still installs everything
It works! Thanks, Oren. On Nov 8, 12:05 pm, Oren Teich o...@heroku.com wrote: Bundlr 1.0.3 changed the syntax for without. We are deploying a hotfix tomorrow. In the meantime, seperate the groups you don't want deployed with a : instead of a space. e.g. heroku config:add BUNDLER_WITHUOUT=development:test Oren On Mon, Nov 8, 2010 at 12:03 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: Yes, this started happening for me on Friday (worked fine up until then). and it looks like it's still going on. When I push I see Using --without development test, but it tries to install all the development/test gems anyway (which it can't, for me, because Heroku can't handle ruby-debug19). I haven't had a burning need to deploy recently, so I've been waiting for Heroku to sort it out, but it looks like they haven't, yet. Anyway, it's definitely a bug, so why not open a support ticket about it? And maybe let us know how it works out? On Nov 6, 3:06 pm, Volkan Unsal spockspla...@gmail.com wrote: I wanted Heroku's bundler to ignore some gems I have installed. It doesn't manage to do that, and as a result I am getting lots of errors. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: http basic auth
Don't see why Heroku would have a problem with it. I've been using Rails' basic auth functionality on Heroku for months with zero issues. On Nov 4, 6:39 pm, Jimmy Thrasher ji...@jimmythrasher.com wrote: Not sure what you mean. If you implement it, it's supported.. it's just a webserver returning a 401 Unauthorized HTTP response. The browser will take care of popping up the dialog. Apache's .htaccess and other non-code ways of accomplishing it aren't supported. Here's an example for Sinatra:http://stackoverflow.com/questions/3479737 Rails has a built-in authenticate_with_http_basic and a couple other methods. Jimmy On Thu, Nov 4, 2010 at 9:09 PM, Brandon Casci bran...@loudcaster.comwrote: Anyone know if basic http auth supported on Heroku? I don't get the login box so I'm assuming no, unless there is a special convention I don't know about. -- = Brandon Casci Loudcaster http://loudcaster.com = -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- +1-919-627-7546 -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Pushing a Rails app without the rails gem.
For anyone curious about this, I eventually solved it by starting off my Gemfile with: source :rubygems # Shut up, Heroku. # gem 'rails' gem 'railties', '3.0.0' gem 'actionpack', '3.0.0' gem 'activesupport', '3.0.0' And so on... On Oct 18, 11:54 am, Chris Hanks christopher.m.ha...@gmail.com wrote: It's a Rails 3 app, so it has an automatically generated config.ru. Thanks for the suggestion, though. On Oct 18, 9:38 am, Jimmy Thrasher ji...@jimmythrasher.com wrote: I'm not sure about this, but you could try setting up a config.ru file to manually start up your Rails app. My suspicion is that it's failing to find config.ru, auto-detecting a Rails app, and not finding all the gems. Just some thoughts.. maybe you've already done all that. :) Jimmy On Mon, Oct 18, 2010 at 12:33 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: There's a lot of parts of Rails that my app has no use for (activerecord, actionmailer, activeresource, and all their associated dependencies like arel, mail, and so on), and I'm trying to take advantage of the new modularity in Rails 3. In my Gemfile I've replaced the rails dependency with: gem 'railties', '3.0.0' gem 'actionpack', '3.0.0' gem 'activesupport', '3.0.0' And my app works totally fine this way, except that Heroku won't accept it, saying: - Rails app detected ! Heroku Bamboo does not include any Rails gems by default. ! You'll need to declare it in either .gems or Gemfile. ! Seehttp://docs.heroku.com/gemsfordetails on specifying gems. ! Heroku push rejected, no Rails gem specified. Is there a workaround for this? Something I can change to not trigger the Rails app detection? I understand the need for the warning, but ideally, I think my invocation of 'railties' would signify to Heroku that I know what I'm doing. Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- +1-919-627-7546 -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: heroku routing issues
Can you explain what is actually happening? Is there an error message? Also, actions aren't typically defined in the ApplicationController, so I wouldn't be surprised if putting them there makes things iffy. You can try running your app in production mode on your local machine is order to see whether the problem is the environment or Heroku (try rails s production at the command line). On Oct 29, 12:33 pm, Cris cristinarand...@gmail.com wrote: I should have also added the routing problems occur in the production environment (on heroku) but not in dev environment. Thanks On Oct 29, 2:15 pm, Cris cristinarand...@gmail.com wrote: Hi there, I'm having problems with routing in my heroku environment heroku. I have this redirect in the index method of the application controller. class ApplicationController ActionController::Base . def index redirect_to root_url end end this works perfectly in my local environment yet in heroku it fails (does not get redirectly). All other routing is functioning correctly, all except for the index method of ApplicationController. Is there something that is causing this in the heroku environment that I have overlooked? Thanks in advance, I'm new to the heroku environment. Thanks, Cris -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: heroku routing issues
My advice would be to drop the index action entirely from ApplicationController, since its not very good practice (ApplicationController isn't meant to define actions - it's more for collecting before_filters and things like that). If you'd like to talk a bit more about what you're trying to do with the index action there, we can certainly give you some advice on what to replace it with. On Oct 29, 2:40 pm, Cris cristinarand...@gmail.com wrote: Hi there, thanks for the reply. The actions are defined in index of ApplicationController. There is no error message, simply that the code within index is not invoked. I ran in production mode locally, and it indeed gives the same error. I checked the routes and they are the same, after running rake routes. Also I'm running rails 2.3.8 It doesn't appear that code should be used in the index method of Action Controller... On Oct 29, 4:00 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: Can you explain what is actually happening? Is there an error message? Also, actions aren't typically defined in the ApplicationController, so I wouldn't be surprised if putting them there makes things iffy. You can try running your app in production mode on your local machine is order to see whether the problem is the environment or Heroku (try rails s production at the command line). On Oct 29, 12:33 pm, Cris cristinarand...@gmail.com wrote: I should have also added the routing problems occur in the production environment (on heroku) but not in dev environment. Thanks On Oct 29, 2:15 pm, Cris cristinarand...@gmail.com wrote: Hi there, I'm having problems with routing in my heroku environment heroku. I have this redirect in the index method of the application controller. class ApplicationController ActionController::Base . def index redirect_to root_url end end this works perfectly in my local environment yet in heroku it fails (does not get redirectly). All other routing is functioning correctly, all except for the index method of ApplicationController. Is there something that is causing this in the heroku environment that I have overlooked? Thanks in advance, I'm new to the heroku environment. Thanks, Cris -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Bundler upgrade to 1.0.3 to support BUNDLE_WITHOUT?
Yep, it's worked for me for weeks too. On Oct 28, 7:28 am, marcel mpoi...@gmail.com wrote: BUNDLE_WITHOUT is currently working for me. It magically started being recognized a few weeks ago. From my heroku config output: BUNDLE_WITHOUT = test development And when I deploy: - Heroku receiving push - Rails app detected - Gemfile detected, running Bundler version 1.0.0 Unresolved dependencies detected; Installing... Using --without test development Fetching source index forhttp://rubygems.org/ ... -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Pushing a Rails app without the rails gem.
It's a Rails 3 app, so it has an automatically generated config.ru. Thanks for the suggestion, though. On Oct 18, 9:38 am, Jimmy Thrasher ji...@jimmythrasher.com wrote: I'm not sure about this, but you could try setting up a config.ru file to manually start up your Rails app. My suspicion is that it's failing to find config.ru, auto-detecting a Rails app, and not finding all the gems. Just some thoughts.. maybe you've already done all that. :) Jimmy On Mon, Oct 18, 2010 at 12:33 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: There's a lot of parts of Rails that my app has no use for (activerecord, actionmailer, activeresource, and all their associated dependencies like arel, mail, and so on), and I'm trying to take advantage of the new modularity in Rails 3. In my Gemfile I've replaced the rails dependency with: gem 'railties', '3.0.0' gem 'actionpack', '3.0.0' gem 'activesupport', '3.0.0' And my app works totally fine this way, except that Heroku won't accept it, saying: - Rails app detected ! Heroku Bamboo does not include any Rails gems by default. ! You'll need to declare it in either .gems or Gemfile. ! Seehttp://docs.heroku.com/gemsfor details on specifying gems. ! Heroku push rejected, no Rails gem specified. Is there a workaround for this? Something I can change to not trigger the Rails app detection? I understand the need for the warning, but ideally, I think my invocation of 'railties' would signify to Heroku that I know what I'm doing. Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- +1-919-627-7546 -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Dalli memcached not hitting the cache
You can use groups with heroku + bundler. For example, I do: heroku config:add BUNDLE_WITHOUT=development test source :rubygems gem 'rails', '3.0.0' group :development, :test do gem 'rspec-rails', '2.0.0' # etc... end Then, Heroku doesn't install any of the RSpec gems. Anyway, it's a bad idea to leave Gemfile.lock out of version control, because then you can't be sure everyone's using the same version of every gem, and you could run into troubles. On Oct 18, 2:52 pm, chris mcclellan...@gmail.com wrote: You can't use groups w/ heroku + bundler. You can do something like if RUBY_PLATFORM =~ /darwin/ ... but then you'll run into problems if you have your Gemfile.lock checked in. To get around this, we just don't check in the Gemfile.lock. To make sure it's working locally, I'd start memcached with memcached - -- that way, you'll be able to see the log directly... which is the best way to determine whether it's your applications fault or heroku's. You should see something like: ree-1.8.7-2010.02 Rails.cache.write(foo, 1) ... and in the memcached log: 8 set foo 0 0 107 NOT FOUND foo 8 STORED ree-1.8.7-2010.02 Rails.cache.read(foo) 8 get foo FOUND KEY foo 8 sending key foo 8 END ree-1.8.7-2010.02 Rails.cache.read(foo-idontexist) = nil 8 get foo-idontexist NOT FOUND foo-idontexist 8 END On Oct 18, 9:40 am, Joost Schuur jsch...@jschuur.com wrote: I've been experimenting with memcached and read good things about the dalli gem, so I thought I'd try it out on Heroku. I can't seem to get it to work and m not getting back any obvious errors either. My production.rb defines: config.action_controller.perform_caching = true config.cache_store = :dalli_store ...and in my Gemfile: gem 'dalli' However, based on the logs, it's not reading from the cache. Upon refreshing the page in question twice (to make sure the cache was written the first time around), I still get this: Exist fragment? views/front_sidebar (506.9ms) Write fragment views/front_sidebar (508.3ms) Likewise, from the heroku console for the app in question: Rails.cache.write('hello', 'world') = false Rails.cache.read('hello') = nil I've restarted the heroku server during my tests too, and have confirmed that I have the add-on installed (the app name is tvgridthing-stage, if a Heroku rep would like to take a look). I don't know what other debug tools there are for dalli. I tried checking the stats, but the instructions athttp://docs.heroku.com/memcache#getting-stats-on-usage obviously don't apply to dalli, which doesn't seem to have a stats command. Locally on my Mac, things seem to work just fine, based on the presence of 'Read fragment views/front_sidebar (0.0ms)' in the logs. I'm running Rails 3.0.1 and dalli 0.10.0 FWIW. Then I thought I'd try the standard memcached instead, so I added this to my Gemfile group :production do gem memcache-client gem 'memcached-northscale', :require = 'memcached' end ...and in production.rb: config.cache_store = :mem_cache_store, Memcached::Rails.new Unfortunately, despite the :production group, when I went to do a bundle install locally first, it tries to compile native extensions under Mac OS X for memcached-northscale, and I get a compile error that's been discussed here too: http://osdir.com/ml/ruby-talk/2010-02/msg01693.html Should I be able to tell bundle what environment to run under, so that it'll ignore that on development and thus bypass any attempt to install memcached-norrthscale? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Dogpile effect with Varnish?
Thanks, Ben! On Oct 13, 2:40 pm, Ben Scofield b...@heroku.com wrote: Sorry for the delay! I finally talked to our Varnish expert, and he confirmed that: 1) our configuration should not impede Varnish's default behavior (re: the first question in this thread), and 2) your app's resource configuration (# of dynos, etc.) doesn't affect how much traffic Varnish can handle for it. Our best estimate for Varnish's capacity for a single cached URL is on the order of 4000 requests/second, sustained. I haven't dug deeply into your other thread yet, Thomas -- I'll take another look at it when I can. Ben On Oct 12, 11:13 am, Thomas Balthazar gro...@suitmymind.com wrote: Hi Ben, Any update about this? Thanks, Thomas. On Tue, Oct 5, 2010 at 8:33 PM, Thomas Balthazar gro...@suitmymind.com wrote: Hello Ben, I just read you were about to talk to the Varnish specialist at Heroku. I would really appreciate if you took the time to help me to find the answer to those 2 unanswered questions about Varnish and caching : http://groups.google.com/group/heroku/browse_thread/thread/8e39658d53... http://groups.google.com/group/heroku/browse_thread/thread/fd23e886c2... Thanks in advance for your help! Thomas. On Tue, Oct 5, 2010 at 7:52 PM, Ben Scofield b...@heroku.com wrote: Not sure why this didn't come through earlier, but: I tried out a few experiments, and it looks like our setup doesn't interfere with this default behavior. I'm going to talk to someone with more intimate knowledge of our Varnish config to confirm that, but so far it looks promising. Ben On Oct 5, 12:00 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: Is anyone from Heroku around that might know how their setup works? On Oct 2, 8:42 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I'm wondering about Heroku's use of Varnish. Suppose I have a page that is expensive to produce (lots of database queries) but can be cached in Varnish. Right after Varnish's copy expires, if it's very popular, I might have a dozen people accessing it simultaneously before the newly created version can be stashed in Varnish. So, based on a thread I found (http://www.gossamer-threads.com/lists/ varnish/misc/14750) it looks like Varnish is smart enough by default to only send that expensive request to my backend once, and serve up the response to all the people waiting for it (to prevent a dogpiling effect). But I know that Heroku has its own configuration for Varnish (with lots of servers in a hash ring), and I was wondering whether it's still set up to do this. Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Compiled Slug size not changing after removing large gems from Gemfile
According to the Heroku docs (http://docs.heroku.com/slug-compiler), the git repository is not included in the slug. I've had the same experience as Stephen - I'd specify gems in my Gemfile that would push the slug size up, and then remove them without any effect to the slug size. Then I could destroy the app, create a new one from the same git repository and wind up with a smaller slug size than I'd had before. On Oct 11, 11:52 pm, Keenan Brock kee...@thebrocks.net wrote: Hi Stephen, If you checked in a gem into git Then you deleted the gem from git. Git still has the gem. It is not showing up for today, but git log will show when you added it and removed it. Guess I do not know if you are checking in your bundle dir and gems into git. 1. So are you saying your repo is 17mb, or your slug is 17mb? I'd imagine that the slug would be smaller on your change to the Gemfile. But you stating that your slug is big suggests that maybe the .git directory is included in the slug? The slug size does affect the speed in which a dyno is started / compiles / restarts. But I agree with Chris, I can't imagine it would slow it down too much. --Keenan On Oct 12, 2010, at 12:48 AM, Chris Hanks wrote: Yeah, this has happened to me before, too. I'm pretty sure the problem is on Heroku's end. I don't worry that much about it, though, since the max slug size is 100 MB. On Oct 11, 8:57 pm, stephen murdoch stephenjamesmurd...@gmail.com wrote: I have encountered a few strange problems with Heroku this evening. My repo size somehow managed to bloat to 17mb so I removed some gems from my Gemfile (ran bundle install, git add Gemfile.lock etc etc) and found that my repo was still 17mb. I removed all my gems (apart from rails and pg) and the repo didn't get any smaller. My app is no bigger than 2mb. So I deleted my app (along with my .git and .bundle directories) and created everything from scratch. This seemed to fix the problem, until just now, when I tried to remove 4 large gems from my Gemfile and didn't see any change in the size of my compiled slug on Heroku. Is this something that other people are experiencing? FWIW, I've tried everything I can think of, including removing Gemfile.lock, .bundle etc etc -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Dogpile effect with Varnish?
Thanks for looking into this stuff, Ben. On Oct 5, 11:33 am, Thomas Balthazar gro...@suitmymind.com wrote: Hello Ben, I just read you were about to talk to the Varnish specialist at Heroku. I would really appreciate if you took the time to help me to find the answer to those 2 unanswered questions about Varnish and caching :http://groups.google.com/group/heroku/browse_thread/thread/8e39658d53...http://groups.google.com/group/heroku/browse_thread/thread/fd23e886c2... Thanks in advance for your help! Thomas. On Tue, Oct 5, 2010 at 7:52 PM, Ben Scofield b...@heroku.com wrote: Not sure why this didn't come through earlier, but: I tried out a few experiments, and it looks like our setup doesn't interfere with this default behavior. I'm going to talk to someone with more intimate knowledge of our Varnish config to confirm that, but so far it looks promising. Ben On Oct 5, 12:00 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: Is anyone from Heroku around that might know how their setup works? On Oct 2, 8:42 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I'm wondering about Heroku's use of Varnish. Suppose I have a page that is expensive to produce (lots of database queries) but can be cached in Varnish. Right after Varnish's copy expires, if it's very popular, I might have a dozen people accessing it simultaneously before the newly created version can be stashed in Varnish. So, based on a thread I found (http://www.gossamer-threads.com/lists/ varnish/misc/14750) it looks like Varnish is smart enough by default to only send that expensive request to my backend once, and serve up the response to all the people waiting for it (to prevent a dogpiling effect). But I know that Heroku has its own configuration for Varnish (with lots of servers in a hash ring), and I was wondering whether it's still set up to do this. Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Dogpile effect with Varnish?
Is anyone from Heroku around that might know how their setup works? On Oct 2, 8:42 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I'm wondering about Heroku's use of Varnish. Suppose I have a page that is expensive to produce (lots of database queries) but can be cached in Varnish. Right after Varnish's copy expires, if it's very popular, I might have a dozen people accessing it simultaneously before the newly created version can be stashed in Varnish. So, based on a thread I found (http://www.gossamer-threads.com/lists/ varnish/misc/14750) it looks like Varnish is smart enough by default to only send that expensive request to my backend once, and serve up the response to all the people waiting for it (to prevent a dogpiling effect). But I know that Heroku has its own configuration for Varnish (with lots of servers in a hash ring), and I was wondering whether it's still set up to do this. Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Dogpile effect with Varnish?
I'm wondering about Heroku's use of Varnish. Suppose I have a page that is expensive to produce (lots of database queries) but can be cached in Varnish. Right after Varnish's copy expires, if it's very popular, I might have a dozen people accessing it simultaneously before the newly created version can be stashed in Varnish. So, based on a thread I found (http://www.gossamer-threads.com/lists/ varnish/misc/14750) it looks like Varnish is smart enough by default to only send that expensive request to my backend once, and serve up the response to all the people waiting for it (to prevent a dogpiling effect). But I know that Heroku has its own configuration for Varnish (with lots of servers in a hash ring), and I was wondering whether it's still set up to do this. Thanks! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Heroku Mongoid config and how it works
Heroku might be breaking up the ENV['MONGOHQ_URL'] into those components for you. Try heroku console from your command line and then ENV['MONGOID_HOST'] to see if it's present. On Sep 27, 6:51 am, Abel Tamayo abel.tam...@gmail.com wrote: It seems MongoMapper is more popular around here. At least it's the solution I'm using and works flawlessly with Heroku. On Mon, Sep 27, 2010 at 2:33 PM, Bradley bradleyrobert...@gmail.com wrote: no one has a comment on this? I thought it'd be a fairly straight forward answer. On Sep 22, 7:11 pm, Bradley bradleyrobert...@gmail.com wrote: I just used a template for a new rails app that sets up everything for use with Mongoid. The mongoid.yml file looks like this: production: host: %= ENV['MONGOID_HOST'] % port: %= ENV['MONGOID_PORT'] % username: %= ENV['MONGOID_USERNAME'] % password: %= ENV['MONGOID_PASSWORD'] % database: %= ENV['MONGOID_DATABASE'] % I added the MONGOHQ_URL config var to my app (with the string given from mongohq), deployed and it worked! My question is... HOW? I don't get how adding that single URL has all of a sudden given me these 5 environment variables. I searched through the heroku docs on mongo and they pretty sparse. Nowhere does it mention that I might use these particular env vars. So how does this happen? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Heroku Mongoid config and how it works
I guess they wanted the MongoHQ addon to just work with Mongoid in addition to MongoMapper. On Sep 27, 4:58 pm, Bradley bradleyrobert...@gmail.com wrote: ya you're right, those env vars exist. i wonder why the docs on this are so sparse, and how the author of this template found this out. I haven't been able to locate anything that gives an indication that this is happening. On Sep 27, 11:44 am, Chris Hanks christopher.m.ha...@gmail.com wrote: Heroku might be breaking up the ENV['MONGOHQ_URL'] into those components for you. Try heroku console from your command line and then ENV['MONGOID_HOST'] to see if it's present. On Sep 27, 6:51 am, Abel Tamayo abel.tam...@gmail.com wrote: It seems MongoMapper is more popular around here. At least it's the solution I'm using and works flawlessly with Heroku. On Mon, Sep 27, 2010 at 2:33 PM, Bradley bradleyrobert...@gmail.com wrote: no one has a comment on this? I thought it'd be a fairly straight forward answer. On Sep 22, 7:11 pm, Bradley bradleyrobert...@gmail.com wrote: I just used a template for a new rails app that sets up everything for use with Mongoid. The mongoid.yml file looks like this: production: host: %= ENV['MONGOID_HOST'] % port: %= ENV['MONGOID_PORT'] % username: %= ENV['MONGOID_USERNAME'] % password: %= ENV['MONGOID_PASSWORD'] % database: %= ENV['MONGOID_DATABASE'] % I added the MONGOHQ_URL config var to my app (with the string given from mongohq), deployed and it worked! My question is... HOW? I don't get how adding that single URL has all of a sudden given me these 5 environment variables. I searched through the heroku docs on mongo and they pretty sparse. Nowhere does it mention that I might use these particular env vars. So how does this happen? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Question about HTTP caching
I've read through the caching docs, and Things Caches Do, but I'm still not clear on something, so here I am. Suppose I signal Heroku to cache a page for an hour (setting the 'Cache-Control' header to 'public, max-age=3600'). Alice visits the page at noon, and Heroku stashes it in Varnish to expire at 1 p.m. It's also cached in Alice's browser until 1 p.m., so that if she requests it again before then, the request won't even reach Heroku. Bob visits the same page at 12:30. Heroku still has it in Varnish, so it returns it from there, and it's cached in Bob's browser. My question is, when will Bob's browser expire it? At 1 p.m. (the time my app told Heroku/Varnish it was good until), or 1:30 (one hour from the time Bob received it)? Thanks -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Question about HTTP caching
Ok, thanks! On Sep 9, 5:09 pm, Pedro Belo pe...@heroku.com wrote: At the time you told Heroku, 1pm. Thanks, Pedro On Thu, Sep 9, 2010 at 4:59 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: I've read through the caching docs, and Things Caches Do, but I'm still not clear on something, so here I am. Suppose I signal Heroku to cache a page for an hour (setting the 'Cache-Control' header to 'public, max-age=3600'). Alice visits the page at noon, and Heroku stashes it in Varnish to expire at 1 p.m. It's also cached in Alice's browser until 1 p.m., so that if she requests it again before then, the request won't even reach Heroku. Bob visits the same page at 12:30. Heroku still has it in Varnish, so it returns it from there, and it's cached in Bob's browser. My question is, when will Bob's browser expire it? At 1 p.m. (the time my app told Heroku/Varnish it was good until), or 1:30 (one hour from the time Bob received it)? Thanks -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Question about HTTP caching
Really? Why doesn't Heroku partition certain URLs to certain servers (through a hashing scheme or something similar)? I'm curious now :) How many servers are there? On Sep 9, 5:18 pm, Pedro Belo pe...@heroku.com wrote: On a side note though, we run multiple caching servers - so there's no guarantee that Bob's request will hit the same server. If you need more control to your cache we recommend using Memcache. On Thu, Sep 9, 2010 at 5:14 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: Ok, thanks! On Sep 9, 5:09 pm, Pedro Belo pe...@heroku.com wrote: At the time you told Heroku, 1pm. Thanks, Pedro On Thu, Sep 9, 2010 at 4:59 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: I've read through the caching docs, and Things Caches Do, but I'm still not clear on something, so here I am. Suppose I signal Heroku to cache a page for an hour (setting the 'Cache-Control' header to 'public, max-age=3600'). Alice visits the page at noon, and Heroku stashes it in Varnish to expire at 1 p.m. It's also cached in Alice's browser until 1 p.m., so that if she requests it again before then, the request won't even reach Heroku. Bob visits the same page at 12:30. Heroku still has it in Varnish, so it returns it from there, and it's cached in Bob's browser. My question is, when will Bob's browser expire it? At 1 p.m. (the time my app told Heroku/Varnish it was good until), or 1:30 (one hour from the time Bob received it)? Thanks -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: free vs. paid heroku app performance
What other questions do you guys have on the area that I should include? Two: What is Heroku's timeout when spinning up dynos/workers? I thought that I'd seen this mentioned somewhere, but I can't find it now. I ask because an app I'm thinking of would need to hit external services and the database when starting up, which could take a while. Is there a limit to how long a daily cron job can run? I'm planning on some number crunching that could take up to an hour or so. Thank you! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Backup for MongoHQ Data
Not sure what that error message means, sorry. You might ask in the MongoDB google group - 10gen is good about helping users with issues: http://groups.google.com/group/mongodb-user Also, I should have mentioned before that there's a guide to mongorestore and the other MongoDB import/export tools here: http://www.mongodb.org/display/DOCS/Import+Export+Tools#ImportExportTools-mongorestore On Aug 26, 2:22 pm, Ginny Hendry cont...@ghendry.com wrote: Chris- Thanks. That got me most of the way there but I needed to extract those four fields from the URL I connect with (in ENV['MONGOHQ_URL']) that looks like this: mongodb://app123456:alongstr...@flame.mongohq.com:27078/app123456 This dump command seemed to work for me: mongodump -hflame.mongohq.com:27078 -dapp123456 -uapp123456 - palongstring It created a dump subdirectory with what looks like my collections. Now how do I restore it to a local database? I tried several variations of this command mongorestore -hlocalhost:27017 -dapp123456 and got: connected to: localhost:27017 don't know what to do with [dump] so I'm not sure what mongorestore wants. I am running mongod 1.6.1 locally with default settings. Thanks. -Ginny On Aug 26, 12:45 am, Chris Hanks christopher.m.ha...@gmail.com wrote: Use the mongodump utility from the command line. It comes with mongodb, and can be pointed at whatever remote database you like. For example, try a ruby script that looks something like: host = 'flame.mongohq.com:27000' db = 'ginnys-database' user = 'ginny' pass = 'password' `mongodump -h#{host} -d#{db} -u#{user} -p#{pass} --out ~/dump` That'll dump the entire contents to your local ~/dump folder. If you need to restore them, you can use mongorestore, which takes similar arguments. On Aug 25, 9:44 pm, Ginny Hendry cont...@ghendry.com wrote: Am I right in assuming that Heroku bundles don't include MongoHQ data? Do db:pull or taps work with MongoHQ? If not, what are our options for backup and restore for MongoHQ databases? I'm sure MongoHQ has backups in case their disks crash but I need my own backups in case I or a user destroys something important. Thx. -Ginny -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: ANNOUNCEMENT: Upgrading to Bundler 1.0.0.rc.6
Hurrah! Is there any word/ETA on Ruby 1.9.2? On Aug 24, 4:16 pm, Terence Lee tere...@heroku.com wrote: Hello, Bundler 1.0.0.rc.6 was pushed out yesterday with some bug fixes. You can view the complete changelog here:http://github.com/carlhuda/bundler/blob/master/CHANGELOG.md The new Rails 3 RC2 also requires the newest version of bundler. We're planning to rollout the new version of bundler Thursday afternoon. Please be ready for the upgrade and remember to use staging apps before pushing code to production. Thanks, Terence -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Backup for MongoHQ Data
Use the mongodump utility from the command line. It comes with mongodb, and can be pointed at whatever remote database you like. For example, try a ruby script that looks something like: host = 'flame.mongohq.com:27000' db = 'ginnys-database' user = 'ginny' pass = 'password' `mongodump -h#{host} -d#{db} -u#{user} -p#{pass} --out ~/dump` That'll dump the entire contents to your local ~/dump folder. If you need to restore them, you can use mongorestore, which takes similar arguments. On Aug 25, 9:44 pm, Ginny Hendry cont...@ghendry.com wrote: Am I right in assuming that Heroku bundles don't include MongoHQ data? Do db:pull or taps work with MongoHQ? If not, what are our options for backup and restore for MongoHQ databases? I'm sure MongoHQ has backups in case their disks crash but I need my own backups in case I or a user destroys something important. Thx. -Ginny -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Daily cron job time - can it be changed?
You can open a support ticket to request a specific time. Chris On Jul 27, 10:37 am, Henry Wagner hjw3...@gmail.com wrote: Hi, I have an app on Heroku with a daily cron job. Currently the job runs at around 11pm. Is this time configurable? Ideally I would like to have the job run at around 9am. Thanks, Henry -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: rails-3.0.0.beta4 on bamboo-ree-1.8.7?
I'm anxious for Bundler 0.9.26 also! On Jun 9, 3:31 am, Ariejan de Vroom arie...@gmail.com wrote: Hey, I'm running into a Bundler/Rails issue when trying to deploy a Rails 3.0.0.beta4 app to the bamboo-ree-.1.8.7 stack. Rails 3b4 need bundler 0.9.26, however, 0.9.25 is installed. Are there any plans on upgrading Bundler on the bamboo stack or are there other ways around? Thanks, Ariejan de Vroomhttp://ariejan.net -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: NOTICE: Memcached exiting beta
Sounds good. I wasn't a member of the beta - was the flush_all function ever added? I don't see it in the docs. Thanks On Apr 19, 5:48 pm, Oren Teich o...@heroku.com wrote: Hi all, We're excited to announce that we'll be taking memcached out of beta this week. We will have a few plans to meet different needs: 5MB - Free 100MB - $20/month 250MB - $40/month 1GB - $90/month 10GB - $800/month 50GB - $3500/month All existing users of the memcached add-on will be grandfathered for 2 weeks. After two weeks, you will be charged at the 100MB plan level ($20/month). Please let me know if you have any questions. Oren -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: heroku restart inside app for clearing http cache
Chap's question is pretty clear - he wants to flush a specific page from the http cache. Unfortunately, there's no way to do this on heroku right now. I'd love this functionality, too, but it just doesn't seem to be possible yet. What you can do, though, is set a low expiration on the cached content. For example, if you use the example shown in the heroku http caching docs: response.headers['Cache-Control'] = 'public, max-age=300' Then whenever one of your users edits the content of a page, you can assure them that their contributions have been recorded, and that the page will be updated sometime in the next five minutes. Or you can do 1 minute, 30 seconds, or whatever interval you like. That'll be the most performant approach, but if you need more control than that you won't be able to use HTTP caching just yet. If you're using rails (you didn't mention whether you are) you'll want to look into action or fragment caching in the guide that Carl linked to. You won't be able to use page caching, though - heroku doesn't support it, and offers http caching as an alternative. On Apr 13, 12:42 pm, Carl Fyffe carl.fy...@gmail.com wrote: That isn't forcing an expire, that is we clean up when you push, just to let you know. Please read this:http://tomayko.com/writings/things-caches-do And then this:http://guides.rubyonrails.org/caching_with_rails.html If you still can't figure it out, then come back. You should have a better understanding and can ask a more pointed question. On Tue, Apr 13, 2010 at 2:32 PM, Chap chapambr...@gmail.com wrote: Thanks for responding Carl, I've been going over the docs and the only way it mentions forcing an expire is deploying: http://docs.heroku.com/http-caching#cache-purge-on-deploy On Apr 13, 2:29 pm, Carl Fyffe carl.fy...@gmail.com wrote: There are much easier ways to expire a cache. The docs that explained how to create the cache more than likely will tell you how to expire it. Start there. On Tue, Apr 13, 2010 at 2:23 PM, Chap chapambr...@gmail.com wrote: Need a button for a client to clear the cached version of a resource. As I understand it, redeploying and potentially heroku restart will cause this to happen. Is it possible for the app to restart itself? I wonder how people are handling this immediate cache expire problem. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Maximum RAM per dyno?
On Mar 9, 10:40 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I'm interested in this too. I have several thousand MongoDB documents that are read-only and frequently accessed, so I figured I'd just cache them in the dyno's memory to speed up requests. So is 300 MB the hard limit for each dyno's RAM, then? I suppose that if it grows beyond that point, the dyno is restarted? Does anyone have an answer for this? Thanks in advance! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Maximum RAM per dyno?
No, I'm familiar with http caching, and it's not what I'm looking to do. Thanks anyway, though. What I'm doing is actually not that complex. MongoMapper already has an identity map, I'll just be tweaking it to persist between requests. And I'm only doing this for a few of my models (ones that are accessed somewhat randomly by id several times per request, and whose records are only modified during site maintenance anyway). It's not like I'm trying to write my own caching system from scratch. Anyway, can someone verify that 300 MB is the maximum RAM available for a dyno? I don't expect to get near it anytime soon, but it would be helpful to know. Thanks! On Mar 10, 11:36 am, Carl Fyffe carl.fy...@gmail.com wrote: Chris, Will this work for you?http://docs.heroku.com/http-caching Carl On Wed, Mar 10, 2010 at 2:22 PM, Chris Hanks christopher.m.ha...@gmail.com wrote: On Mar 9, 10:40 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: I'm interested in this too. I have several thousand MongoDB documents that are read-only and frequently accessed, so I figured I'd just cache them in the dyno's memory to speed up requests. So is 300 MB the hard limit for each dyno's RAM, then? I suppose that if it grows beyond that point, the dyno is restarted? Does anyone have an answer for this? Thanks in advance! -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Maximum RAM per dyno?
I'm interested in this too. I have several thousand MongoDB documents that are read-only and frequently accessed, so I figured I'd just cache them in the dyno's memory to speed up requests. So is 300 MB the hard limit for each dyno's RAM, then? I suppose that if it grows beyond that point, the dyno is restarted? On Mar 9, 9:49 pm, Oren Teich o...@heroku.com wrote: I don't have a good answer for your question, but note there is no guarantee that your requests will be served from the same physical machine - we'll move the dyno around as demanded by the cloud. memcached is the way to do persistent (beyond single request) caching. We're a few days away from making it public beta. Oren On Tue, Mar 9, 2010 at 9:02 PM, Alex Chaffee ale...@gmail.com wrote: I've got some frequently-accessed data I'd like to store in RAM between requests. I'm using Sinatra so I'll probably just use an LRU cache in a @@class variable. I think I can muddle through all the technical issues but one: How big can I reasonably make my cache? I.e. how high (or low) should I put my threshold before I start expiring unused data? The only guidance I could find from a quick perusal of heroku.com was onhttp://legal.heroku.com/aup:Dyno RAM usage: 300MB - Hard -- which is good to know, but not a complete answer. I'll obviously want to set my cache well below that limit. But without monitoring tools I don't have any idea how much RAM is used by the normal processing of Rack + Sinatra per request, nor do I know how many requests are being serviced per second. My cache is supposed to increase performance, not decrease it by hammering the dyno into swap space, or otherwise interfering with other system functions on the dyno. So... any ideas? Has anyone else done this? Are there any low-level monitoring tools I can use to find out how much RAM I'm currently using, or how loaded the system is, or anything of that nature? Would New Relic help here (and does it work for Sinatra apps)? BTW, although I may want to use memcached as an *additional* caching layer, what I'm interested in exploring now is the feasibility of storing transient data in the app server itself. (I don't want the overhead of instantiating Ruby objects, especially ActiveRecord objects, not to mention that memcached isn't officially available as an addon.) --- Alex Chaffee - a...@cohuman.com -http://alexch.github.com Stalk me:http://friendfeed.com/alexch|http://twitter.com/alexch| http://alexch.tumblr.com -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Pricing of Dynos...
Yes, dynos are billed hourly, but I think he was asking whether he gets billed if the dynos sit there unused. The answer is yes - if you set the dynos to 8, you'll be paying 35 cents an hour regardless of how much traffic they're actually processing. On Jan 16, 12:24 pm, Terence Lee hon...@gmail.com wrote: The latter, once you setup a dyno you're getting build hourly for it. On Sat, 2010-01-16 at 10:29 -0800, holden wrote: I have a newb question, but I'm a bit confused on the pricing since it states on the pricing page estimated monthly cost. I'm not sure how to phrase my question except as an example: If I start a new app, and say I allocated 8 dynos for it, because I'm ambitious that I'm sitting on the next big thing. But lets say my app dives and everyone who uses it hates it and leaves immediately, and really I could have gotten by with 2 dynos. Once I crank the dynos are they constantly up and getting billed? Or are they smart little dynos that only spring into action as the need arises? Thanks for your patience. holden -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Why is it mandatory to submit Credit Card Information for Add Ons ?
Absolutely. I'd much rather trust my credit card number to Heroku than to Paypal. On Jan 14, 4:57 pm, Shane Becker veganstraighte...@gmail.com wrote: Haha, well, let's see if I can rephrase that. Heroku, can you please replace your current payment method with PayPal? It would make things a lot easier for everyone, especially people outside of the US. Dealing with paypal (on either end) has been known from time to time to be the opposite of easy. I have horror stories as both a customer and a developer. At the very very most, they could ADD paypal as a payment option. Certainly, they shouldn't take away the rest. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Filesystem?
My app info screen on heroku.com says: Slug Size: 6.2 MB of 20mb. Is that a typo, then? Should it be 50mb? On Dec 19, 12:25 pm, Oren Teich o...@heroku.com wrote: That is just for the DB. Your slug can be up to 50MB. If you're hosting large assets, we suggest using S3 directly. 1GB is only 0.10/month, and it's blazingly fast. Oren On Sat, Dec 19, 2009 at 6:38 AM, iamjediknight iamjedikni...@gmail.comwrote: I am looking at the Crane package which has 500MB storage. Is that 500MB just for the database or is it 500MB combined for my app (whats in my git repo) and database? Thanks, Scott. On Dec 19, 12:50 am, Adam Wiggins a...@heroku.com wrote: On Fri, Dec 18, 2009 at 7:17 AM, iamjediknight iamjedikni...@gmail.com wrote: I have a rails app that reads liquid templates off the filesystem, as well as content (i.e. images, html, etc.). So will I need to store those via Amazon S3? Is there another way to store content? Assuming these things (templates, images) are in your Git repository and you're only reading them, that will work fine. The only time you can't use the filesystem is for writes, such as storing images or pdfs uploaded by your users. For that you should use S3. Adam -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.comheroku%2bunsubscr...@googlegroups.com . For more options, visit this group at http://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: If you reserve full instance for custom SSL - why don't I get more dynos?
Wojciech, if you ask support about that and get some good news, would you report back? I'm curious about this too. Thanks! Chris On Dec 8, 2:05 pm, Oren Teich o...@heroku.com wrote: I don't know if that's possible or not it's probably a function of the SSL protocol and our routing mesh, but it's beyond my technical knowledge. Best bet is to drop support@ a line, and see what they say. They'll be able to dig into the details for you. Oren On Tue, Dec 8, 2009 at 12:42 PM, Wojciech Kruszewski wojci...@oxos.pl wrote: Thanks Oren, this makes sense. So can that one mostly idle server handle SSL requests for multiple applications? I mean I tried Heroku and was very happy with the experience - looks like it needs little to no maintenance on my part. I'd wish to host a handful smaller web apps, each with 1-3 dynos. I could live with piggyback ssl, if it was my own wildcard certificate. - Wojciech On Dec 8, 8:58 pm, Oren Teich o...@heroku.com wrote: They are totally independent. The way our architecture works, dynos run on machines called railguns, which are specially set up for the job. We have to setup a special (and yes, mostly idle) server just to handle the SSL requests. It's not possible with the product we have today to run dynos on that server. Oren On Tue, Dec 8, 2009 at 7:48 AM, Wojciech Kruszewski wojci...@oxos.pl wrote: Hi, I've read your explanation about why you charge $100/mo for custom SSL (http://docs.heroku.com/ssl#faq). You need exclusive IP, Amazon assigns only one IP for an instance, so you need to reserve full instance just to use one SSL cert - seems fair. Ok, but if you reserve full EC2 instance just for me... then why do I have to pay for extra dynos? Aren't you double-billing for this instance? I believe it's just against your architecture but still I'd like to know the explanation. Regards, Wojciech -- http://twitter.com/WojciechKhttp://oxos.pl-Ruby on Rails development -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Rails 3
Rails 3 won't run on Ruby 1.8.6, so Heroku will need to offer Ruby 1.8.7 or 1.9 (preferably 1.9) before we can use it. This is what's keeping me from trying out Rails 3.0.pre on Heroku right now. Source: http://www.mail-archive.com/rubyonrails-c...@googlegroups.com/msg09683.html On Nov 25, 10:05 am, Pedro Belo pe...@heroku.com wrote: We don't run any customized version of Rails, so most probably yes - you should be able to just run it. On Tue, Nov 24, 2009 at 4:59 PM, jasonb jason.bower...@gmail.com wrote: Just wondering whether people will have the option to play with the 3 beta (once released) on Heroku? -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group athttp://groups.google.com/group/heroku?hl=en. -- You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to her...@googlegroups.com. To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
Re: Database alternative to Postgres?
Can anyone at Heroku comment on this? Thanks! On Oct 21, 3:06 pm, Chris Hanks christopher.m.ha...@gmail.com wrote: Hi - Does Heroku have any plans to offer a NoSQL alternative to Postgres? MongoDB, CouchDB, etc? Thank you! --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Heroku group. To post to this group, send email to heroku@googlegroups.com To unsubscribe from this group, send email to heroku+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/heroku?hl=en -~--~~~~--~~--~--~---